Around 400 scientists, including AI pioneers Geoffrey Hinton and Yoshua Bengio, called on Tuesday for the establishment of an international body to oversee containment standards and report incidents related to advanced AI systems capable of launching cyberattacks.
According to Reuters, researchers stated in an open letter that "advanced AI models should initially be isolated from external networks and subjected to independent oversight," justifying this by arguing that voluntary commitments from the private sector are insufficient.
Containment standards and incident log
The proposed body is expected to set industry-wide containment standards, in addition to maintaining a log of significant security incidents and making them publicly available.
This proposal comes a month after Anthropic, OpenAI, Google, Meta, xAI, and Nvidia signed the White House agreement, under which they committed to taking voluntary safety measures without any formal enforcement mechanism.
Oversteps and breaches
The researchers added that a series of incidents revealed weaknesses in how advanced AI systems are tested and contained, including cases where AI models exceeded isolated testing environments but insufficiently, leading to breaches of external systems.
Signatories from 34 countries stated that "organizations developing or testing models with capabilities to launch cyberattacks must immediately adopt guaranteed containment standards, while governments work on establishing a more formal oversight system."
International controversy and US developments
Debate is growing among policymakers regarding how to oversee advanced AI systems. Henna Verkonen, Executive Deputy for Technological Sovereignty, Security, and Democracy at the European Union, recently told Reuters that the bloc's AI Act is sufficient to address risks posed by rogue AI systems.
The administration of US President Donald Trump lifted the ban on US laboratories two days ago, allowing the British AI Safety Institute early access to models to conduct safety tests.
Trump appointed National Intelligence Director Jay Clayton on October 4 as the official in charge of the AI file in the US administration, a move aimed at enhancing US capabilities in this field and confronting risks associated with its rapid development.
Source: Syrian Arab News Agency (SANA)
Comments (0)